HookThe best marketing campaign of 2019 cost Greggs almost nothing
On 3 January 2019 Greggs launched a vegan sausage roll, developed with Quorn after a petition of more than 20,000 signatures asked for one. It was priced at about £1 — deliberately level with the meat version — and shipped in packaging that parodied a smartphone launch, complete with a slick unboxing video. Then Piers Morgan tweeted his disgust to his millions of followers, calling Greggs 'PC-ravaged clowns', and the Greggs social team replied: 'oh hello Piers, we've been expecting you.' The exchange went viral. Queues formed, shops sold out by mid-morning, like-for-like sales were up nearly 10% within seven weeks, and Greggs raised its profit forecast more than once that year. Paid advertising involved: almost none.
Nothing about that fortnight was luck, and every leaf of 3.5 is inside it. Greggs had understood a customer need (3.5.1) that was never really about vegans: the growth market was flexitarians — meat-eaters cutting down — which is a segmentation insight (3.5.2). The petition and the visible rise of Veganuary were market research it barely had to pay for (3.5.3). And the launch was a marketing mix working as one machine (3.5.4): a product engineered to taste like the original, a price set to remove any excuse not to try it, promotion through earned publicity rather than paid space, and place — hundreds of shops already standing on the walk to work, stocked and ready. Marketing is that whole system pointed at one customer. It is not the adverts bolted on at the end.
ModelWhat customers are actually buying
Every purchase bundles four things customers weigh: price (is it worth it, and does it fit my budget?), quality (does it work, last, taste right?), choice (does a version exist for me?) and convenience (how much time and effort does buying it cost?). Different customers rank them differently — and the same customer ranks them differently at 8am on a station platform than on a Saturday afternoon, which is why Greggs' pavement-to-till speed is a designed feature, not an accident.
Understanding needs is not politeness; it is survival arithmetic. Get it right and sales, repeat purchase and word of mouth follow. Get it wrong and you carry stock nobody wants, spend promotion budget shouting the wrong message, and hand rivals your customers. Blockbuster understood film rental very well and still died, because the underlying need was 'an evening's entertainment with zero effort' — and streaming met it better. That is the second lesson: businesses that merely react to needs are already behind the ones that anticipate them. Greggs did not wait for vegan ranges to become normal on the high street; it moved while the trend was still a rising curve rather than a crowded plateau.
ModelSegmentation: aim before you fire
A market is not a blob; it is groups of people buying for different reasons. Segmentation splits customers by age, gender, income, location and lifestyle so a business can pick a target and tune everything — product, price, message, channel — to that group. Saga aims squarely at the over-50s; energy drinks chase teenagers and shift workers; a farm shop serves the postcodes around it.
Why bother? Focus makes money go further: a targeted product beats a compromise designed for everyone and loved by no one; promotion lands where the target actually looks; and unserved segments become visible gaps in the market — which is exactly how the vegan sausage roll happened. The clever part was defining the segment by behaviour, not by label. Committed vegans were a small minority of the UK population — far too few on their own to justify a national bakery rollout. But flexitarians, people cutting down on meat a few days a week, were a mass market wearing an invisible badge. Aim at them and the design brief writes itself: the roll must sit next to the original, taste nearly identical and cost the same, so that choosing it requires no sacrifice at all. The risks run the other way: segments built on lazy stereotypes misfire, real people belong to several segments at once, and segments move — yesterday's teenage energy-drink buyer is tomorrow's parent reading ingredient labels.
DataResearch: asking, counting, and not fooling yourself
Primary research is new data collected first-hand — questionnaires, interviews, focus groups, observation, product trials. It answers your exact question and stays private, but it is slow, costly and only as good as the sample you asked. Secondary research already exists — government statistics, published market reports, competitors' accounts, review sites and social media chatter. It is fast and cheap, but generic, sometimes stale, and available to your rivals too. Cutting across both: quantitative data counts things (how many, how often, what share) and qualitative data explains them (why, and how it feels). A 200-response questionnaire tells you 62% would try a vegan option; a focus group warns you it had better not taste 'worthy'.
Reliability is the examiner's favourite trap. A small or unrepresentative sample — fifty sixth-formers surveyed outside one shop — can point a business confidently at the wrong answer, and leading questions bake bias into the results before anyone counts them. Once data exists, 3.5.3 expects you to handle it without flinching: calculate percentages and percentage changes, read charts without over-claiming, and compute market share = a firm's sales ÷ total market sales × 100.
A milkshake bar takes £180,000 in a town where total milkshake-and-dessert spending is £1.2 million: market share = 180,000 ÷ 1,200,000 × 100 = 15%. A year later its revenue has grown 15% to £207,000 — but the market has grown to £1.5 million, so share = 207,000 ÷ 1,500,000 × 100 = 13.8%. Sales up, share DOWN: rivals are growing faster, and 'growing but losing share' is exactly the interpretation sentence that turns a 2-mark calculation into a 4-mark answer. Now the reliability caveat: the owner's survey of 50 students outside the door found 70% wanted a new flavour. Fifty self-selected teenagers are not the town. A 400-person sample across age groups might have shown the market's growth was coming from families — and sent the product decision somewhere else entirely.
ModelProduct and price: the first half of the mix
Product decisions cover function, design, quality and range — and whatever makes yours different from the shelf-neighbour: its differentiation, or unique selling point. Products also age. The product life cycle runs introduction → growth → maturity → decline, and each stage demands different behaviour: heavy promotion at introduction, capacity and distribution in growth, price discipline as maturity crowds the market. Extension strategies postpone decline — new flavours, new packaging, new users, new markets. Think of a chocolate bar spawning ice-cream and biscuit versions, or Greggs following the sausage roll with a vegan steak bake to keep the momentum compounding.
Price is a message as well as a number. Cost-plus pricing adds a markup to unit cost: a roll costing 40p to make, marked up 150%, sells at £1. Competitive pricing matches the going rate where products are similar. Penetration pricing launches deliberately low to build share fast, then edges upwards — the streaming-service playbook. Skimming launches high to harvest the keenest early adopters — new consoles, flagship phones — then steps down towards the mass market. A loss leader is priced at or below cost to pull shoppers in to buy profitable items alongside. Greggs' choice was the quietest and smartest of all: parity. At the same £1 as the original, trying the vegan version required no decision, no sacrifice and no maths from the customer.
MechanismPromotion and place: the second half
Promotion is the toolkit for being noticed: paid advertising (TV, outdoor, search and social ads), public relations — coverage you earn rather than buy, which is what the Piers Morgan exchange delivered for nothing — sales promotions (discounts, loyalty schemes, free samples), sponsorship, and social media, where the currency is shareability. The choice is disciplined by two questions: where does the target segment actually look, and what does the budget allow? A national TV slot is wasted on a business with three postcodes of customers; a local café's Instagram, loyalty card and word-of-mouth reputation can outperform it for a thousandth of the cost.
Place is how the product physically reaches the buyer: shops, e-commerce, online marketplaces, apps, click-and-collect, delivery partners. E-commerce hands a small brand national reach with no shopfront; physical presence delivers immediacy and impulse — nobody plans a sausage roll at 8.40am, which is why Greggs' hundreds of high-footfall shops are its distribution strategy, now extended through its app and delivery tie-ups. The through-line of 3.5.4 is coherence: the four Ps must tell one story. A premium product with a bargain price in a discount channel confuses everyone — and the mix is never finished, because the life-cycle stage, the technology and the segment's habits all keep moving.
VocabularyKey terms the mark scheme pays for
TrapsMisconceptions that cost marks
ExamWhat examiners want
Marketing sits on Paper 2, and its data questions are where AQA hides the easiest marks: when a table or chart appears, quote a figure, transform it (a percentage, a percentage change, a market share) and then write one sentence on what it means for this business — the interpretation sentence is the one candidates skip. Set out every calculation as formula → substitution → answer with its unit, because method marks survive arithmetic slips.
On 6-mark 'analyse' questions, chain a mix decision all the way to revenue: parity pricing → no barrier to trial → high first-week volume → queues and social proof → repeat purchase → revenue and market share. On 9-markers you will usually be asked which P to change or which segment to chase: argue two options through the case's constraints — budget, brand image, the segment's habits — then recommend, naming the deciding factor. The classic self-inflicted wound is the floating answer: 'advertise on social media because everyone is online' earns nothing until it names the segment, the platform and why that pairing fits this firm. And never argue past the data: if the case says the survey sampled 30 people, that sample size IS your evaluation point.